Quick Introduction to Artificial Intelligence - AI

 

Artificial Intelligence

Artificial Intelligence is a study and scientific process of creating a digital brain by instructing computers applying different methods without human involvement. In Artificial Intelligence devices learn themselves on the basis of past experiences and learnings.

 

Classical AI was exist some decades ago in which computers perform intelligent tasks only when specific rules are set and an input is given to them upon which they give results. That technology was not enough intelligent, usable and need of Advance AI rose.

Artificial Intelligence is a study and scientific process of creating a digital brain by instructing computers applying different methods without human involvement. In Artificial Intelligence devices learn themselves on the basis of past experiences and learnings.

Now a days AI is Advanced and more intelligent. AI is count as branch of Computer Science in which computers uses methods (Machine and Deep learning) and creating rule from problems and their solutions themselves to perform tasks without taking dictation from humans. 

• AI Value Creation by 2030

The McKinsey Global Institute has predicted substantial value creation of $13 trillion through AI by 2030. This projection is based on AI's potential to revolutionize numerous sectors, such as healthcare, manufacturing, agriculture, automotive, retail and finance. AI is expected to contribute significantly to economic growth and efficiency in various industries, creating value in the trillions of dollars.

• Types of AI

There are three types of Artificial Intelligence on the basis of capabilities.

1. Artificial Narrow or Shallow Intelligence

2. Artificial General Intelligence

3. Artificial Super intelligence

• Artificial Narrow Intelligence:

ANI is also referred as Artificial Shallow Intelligence. In this field, computers are getting specific knowledge through data to perform different task. For example, ChatGPT and Google Bard are recent practical implemention of ANI. 

This type is emerging now a days and people are interested to learn this technology to excel in AI. 

• Artificial General Intelligence and Artificial Super intelligence :

AGI and ASI are types in which computer will be able to perform tasks just like humans. For example thinking, learning and creativity etc. These will also be used to perform multiple tasks, but both types do not exist in today’s world but there are speculations that AGI is going to happen in about next 30 years. 

Popular techniques to do AI are Machine learning and Deep learning?

• Machine learning:

In Machine Learning, there is no need to instruct computers, data is given to the system and according to past experience and data, rules are made to by computers and apply further to get required results.

• Deep learning:

Deep Learning is inspired by biological neurons and study of neural networks is called deep learning. DL is the study of doing ML with neural networks. In 2012, Neural Networks made up of basic and fundamental mathematical functions like comparing two value, equations etc.

• Relation between AI, ML and DL

Deep Learning is a part of Machine Learning and ML exists in domain of AI. 

• Most common used term in AI - Algorithms:

Algorithms are like step-by-step instructions or rules that tell a computer what to do. They help solve problems or perform tasks by breaking them down into smaller, manageable steps. Just like a recipe guides you in cooking, algorithms guide computers in processing information and making decisions.

• Data and it's types in Artificial Intelligence

Data is define as information or facts that are collected, stored, and processed by computers. It can be in various forms such as numbers, text, images, or even sounds. Data is the raw material that computers use to perform tasks, make calculations, and generate insights. It's essentially the building block of information in the digital world.

There are four types of data in AI:

1. Labelled Data:

Any type information with nominated names, is called labelled data. For example, in different house data everything has their specific name like cat, door, floor, TV etc. 

2. Unlabelled Data:

Data in raw form without their nominated names, is called unlabeled data. For example, unarranged vegetables in a basket or text in word document which are not categorized properly. 

3. Structured data:

Data or information in the form of a table, is called structured data. For example, data arranged in excel file with proper categories like Name field, Age etc. 

4. Unstructured Data:

All information which is not in the form of a table, is called unstructured data. It can be in the form of text, image, audio or video etc.

• Three Types of ANI w.r.t functionality

1. Supervised learning:

Supervised learning is a category in which models are trained using labeled data. In other words, the input data is paired with corresponding output labels. A classic example is email spam detection, where the algorithm learns to classify emails as spam or not spam based on labeled training data. In this learning method, the computer is taught by using labelled data (telling names of things) several times, then asking the computer,whether it is a particular thing or not, and the computer will give the correct answer.

AI Supervised Learning deals with two mode

● Training mode (for updation)

● Inference mode (for Prediction)

If training is happening, the system will update but there will be no prediction and vice versa.

Both modes can not be perform together.

2. Unsupervised Learning:

Unsupervised learning involves training models on unlabeled data. The system identifies patterns and structures within the data without predefined categories. A common application is clustering, where the system groups similar data points together, as seen in customer segmentation for marketing purposes.

In unsupervised learning, unlabelled data is used.In this learning, unlabelled data is given to the computer and then the system will make groups according to different attributes like same shape, size,colour etc.

3. Reinforcement Learning:

Reinforcement learning is an approach where AI agents learn to interact with their environment by receiving feedback or rewards based on their actions. This type of learning is commonly employed in training autonomous systems, like self-driving cars, which learn to navigate safely by receiving feedback through sensors

This learning is based on trial and error method, system is given a specific task, it will perform it once, then again same task will be given to the system, it will perform it in a better way than previous, after several times, system will be able to find out the most optimized way to do that task. This learning method is known as reinforcement.

All we have discussed above is normally related to Discriminative AI

Discriminative AI 

It is type of artificial intelligence that focuses on learning to distinguish between different categories or classes in data. It's often used in tasks such as classification, where the AI learns to identify patterns or features that differentiate one class from another. Unlike generative AI, which aims to create new data similar to existing examples, discriminative AI aims to make decisions based on existing data patterns.

Read more blogs from Abdullah:

Deep and Machine Learning

Discriminative and Generative AI

Prompt Engineering 

Diffusion Model in AI